Segment 1: The Basics of ggplot2

The `{ggplot2}` Package

{ggplot2} is a system for declaratively creating graphics,
based on “The Grammar of Graphics” (Wilkinson, 2005). You provide the data, tell {ggplot2} how to map variables to aesthetics, what graphical primitives to use, and it takes care of the details.

Advantages of `{ggplot2}`

consistent underlying grammar of graphics (Wilkinson, 2005)
very flexible, layered plot specification
theme system for polishing plot appearance
active and helpful community

The Setup

The Course Material

Clone the repository.

Open the R Project: hands-on-ggplot2.Rproj

Open the first script: 01-grammar.Rmd

Run code locally

… or knit the report

The Package

ggplot2 is a data visualization package for the programming language R created by Hadley Wickham.

It should be already installed on your system (if not run the first line in the following chunk). The functionality of the package can be loaded by calling library() as for any other package:

#install.packages("ggplot2")
library(ggplot2)

ggplot2 is part of the tidyverse package collection. Thus, you can also load tidyverse without running library(ggplot2):

#install.packages("tidyverse")
library(tidyverse)

The Data Set

We use cryptocurrency financial data, pulled from CoinMarketCap.com. For our purposes, we limit the data to the period 08/2017–12/2019 and the top 4 cryptocurrencies.

I have already prepared the data. If you want to know how, you can have a look here.

Import the Data

Using the read_csv() function from the {readr} package, we can read the data directly from the web:

url <- "https://raw.githubusercontent.com/z3tt/hands-on-ggplot2/main/data/crypto_cleaned.csv"

data <- readr::read_csv(url)

data

# A tibble: 2,812 x 9
   currency     date        open  high   low close  year month  yday
   <chr>        <date>     <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
 1 binance-coin 2019-12-04  15.4  15.7  15.0  15.3  2019    12   338
 2 binance-coin 2019-12-03  15.2  15.6  15.0  15.3  2019    12   337
 3 binance-coin 2019-12-02  15.5  15.7  15.2  15.2  2019    12   336
 4 binance-coin 2019-12-01  15.7  15.7  15.0  15.5  2019    12   335
 5 binance-coin 2019-11-30  16.3  16.4  15.5  15.7  2019    11   334
 6 binance-coin 2019-11-29  15.7  16.3  15.6  16.3  2019    11   333
 7 binance-coin 2019-11-28  16.1  16.2  15.6  15.7  2019    11   332
 8 binance-coin 2019-11-27  15.5  16.2  14.9  16.1  2019    11   331
 9 binance-coin 2019-11-26  15.3  15.9  15.2  15.5  2019    11   330
10 binance-coin 2019-11-25  15.3  15.7  14.2  15.3  2019    11   329
# ... with 2,802 more rows

Of course, one can import local files as well:

data_local <- readr::read_csv("data/crypto_cleaned.csv")

data_local

# A tibble: 2,812 x 9
   currency     date        open  high   low close  year month  yday
   <chr>        <date>     <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
 1 binance-coin 2019-12-04  15.4  15.7  15.0  15.3  2019    12   338
 2 binance-coin 2019-12-03  15.2  15.6  15.0  15.3  2019    12   337
 3 binance-coin 2019-12-02  15.5  15.7  15.2  15.2  2019    12   336
 4 binance-coin 2019-12-01  15.7  15.7  15.0  15.5  2019    12   335
 5 binance-coin 2019-11-30  16.3  16.4  15.5  15.7  2019    11   334
 6 binance-coin 2019-11-29  15.7  16.3  15.6  16.3  2019    11   333
 7 binance-coin 2019-11-28  16.1  16.2  15.6  15.7  2019    11   332
 8 binance-coin 2019-11-27  15.5  16.2  14.9  16.1  2019    11   331
 9 binance-coin 2019-11-26  15.3  15.9  15.2  15.5  2019    11   330
10 binance-coin 2019-11-25  15.3  15.7  14.2  15.3  2019    11   329
# ... with 2,802 more rows

This assumes that you have placed the file in a folder called data in your working directory.
You can specify this directory via setwd() or, and preferably, use R projects.

Short Explanation of ::

The so–called namespace allows to access functions from a package directly without loading it first.

packagename::function(argument)

Furthermore, it helps readers to understand from which package a function is imported from.

Exercise 1

Import the data set on Chicago weather data:
https://raw.githubusercontent.com/z3tt/ggplot-courses/master/data/chicago-nmmaps.csv
Inspect the data.

The Grammar of Graphics

The Structure of ggplot2

Data: ggplot(data)

We need to specify the data in the ggplot() call:

ggplot(data = data)

There is only an empty panel because ggplot2 doesn’t know what of the data it should plot.

Aesthetics: aes()

We need to specify two variables we want to plot as positional aesthetics:

ggplot(data = data, mapping = aes(x = date, y = close))

There is only an empty panel because ggplot2 doesn’t know how it should plot the data.

Thanks to implicit matching of arguments in ggplot() and aes(), we can also write:

ggplot(data, aes(date, close))

Layers: geom_() and stat_()

By adding one or multiple layers we can tell ggplot2 how to represent the data. There are lots of built-in geometric elements (geom's) and statistical transformations (stat's):

Adapted from ggplot2.tidyverse.org/reference

We can tell ggplot2 to represent the data for example as a scatter plot:

ggplot(data, aes(date, close)) +
  geom_point()

Aesthetics: aes() (again)

Aesthetics do not only refer to x and y positions, but also groupings, colors, fills, shapes etc.

ggplot(data = data, mapping = aes(x = date, y = close, color = currency)) +
  geom_point()

Important Things First: Change the Default Theme

You can replace the default theme with one of the other built-in themes with theme_set(). Note that you can as well adjust some global settings, for example the base_size which is often too small in the default (11).

theme_set(theme_light(base_size = 18))

By using theme_set() the new theme is used for any plot you create aftwerwards! Give it a try on go back to the last chunk and re-run the code to generate the colored scatter plot.

Layers: geom_() and stat_() (repeated)

The exciting thing about layers is that you can combine several geom_*() and stat_*() calls:

ggplot(data, aes(date, close, color = currency)) +
  geom_line() +
  geom_point()

… and aesthetics can be applied either globally:

ggplot(data, aes(date, close, color = currency, shape = currency)) +
  geom_line() +
  geom_point()

… or for each layer individually:

ggplot(data, aes(date, close)) +
  geom_line(aes(color = currency)) +
  geom_point(aes(shape = currency))

Exercise 2

If needed, import the Chicago weather data again:

chic <- readr::read_csv(
  "https://raw.githubusercontent.com/z3tt/ggplot-courses/master/data/chicago-nmmaps.csv"
)

Create a scatter plot of temperature (temp) versus day (date).
Color the points by season (season).
Color the points by year (year).
- What’s the problem? How could you fix it?
Bonus: Turn the scatter plot, colored by season, into a line plot.
- What’s the problem? How could you fix it?

Saving a `ggplot`

You can export your plot via the ggsave() function:

ggsave(filename = "my_ggplot.pdf", 
       width = 10, height = 7, 
       device = cairo_pdf)

ggsave(filename = "my_ggplot.png", 
       width = 10, height = 7, 
       dpi = 700)

-> Scales, Coordinate Systems, Facets, Themes, and Annotations will follow later

Resources

“ggplot2: Elegant Graphics for Data Analysis”, free–access book by Hadley Wickham et al.
{ggplot2} reference
{ggplot2} cheatsheet
“R for Data Science”, free–access book by Hadley Wickham
“Data Visualization: A Practical Introduction”, free–access book by Kieran Healy
“A {ggplot2} Tutorial for Beautiful Plotting in R”, my extensive “how to”-tutorial

Appendix

Import the Data

The `{here}` Package

A good workflow when working with local files is offered by the {here} package in combination with R projects:

here::here()

[1] "C:/Users/DataVizard/Google Drive/Work/DataViz/Teaching/2022_OReilly_Trainings/hands-on-ggplot2-training"

data_local <- readr::read_csv(here::here("data", "crypto_cleaned.csv"))

data_local

# A tibble: 2,812 x 9
   currency     date        open  high   low close  year month  yday
   <chr>        <date>     <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
 1 binance-coin 2019-12-04  15.4  15.7  15.0  15.3  2019    12   338
 2 binance-coin 2019-12-03  15.2  15.6  15.0  15.3  2019    12   337
 3 binance-coin 2019-12-02  15.5  15.7  15.2  15.2  2019    12   336
 4 binance-coin 2019-12-01  15.7  15.7  15.0  15.5  2019    12   335
 5 binance-coin 2019-11-30  16.3  16.4  15.5  15.7  2019    11   334
 6 binance-coin 2019-11-29  15.7  16.3  15.6  16.3  2019    11   333
 7 binance-coin 2019-11-28  16.1  16.2  15.6  15.7  2019    11   332
 8 binance-coin 2019-11-27  15.5  16.2  14.9  16.1  2019    11   331
 9 binance-coin 2019-11-26  15.3  15.9  15.2  15.5  2019    11   330
10 binance-coin 2019-11-25  15.3  15.7  14.2  15.3  2019    11   329
# ... with 2,802 more rows

Base R

The base R function read.csv() works in the same way as readr::read_csv():

data <- read.csv(here::here("data", "crypto_cleaned.csv"))

head(data)

      currency       date  open  high   low close year month yday
1 binance-coin 2019-12-04 15.35 15.69 15.01 15.28 2019    12  338
2 binance-coin 2019-12-03 15.19 15.55 15.05 15.31 2019    12  337
3 binance-coin 2019-12-02 15.51 15.71 15.15 15.19 2019    12  336
4 binance-coin 2019-12-01 15.74 15.74 15.05 15.50 2019    12  335
5 binance-coin 2019-11-30 16.26 16.37 15.54 15.72 2019    11  334
6 binance-coin 2019-11-29 15.68 16.34 15.65 16.27 2019    11  333

… and we can turn it into a tibble afterwards:

data <- tibble::as_tibble(data)

data

# A tibble: 2,812 x 9
   currency     date        open  high   low close  year month  yday
   <chr>        <chr>      <dbl> <dbl> <dbl> <dbl> <int> <int> <int>
 1 binance-coin 2019-12-04  15.4  15.7  15.0  15.3  2019    12   338
 2 binance-coin 2019-12-03  15.2  15.6  15.0  15.3  2019    12   337
 3 binance-coin 2019-12-02  15.5  15.7  15.2  15.2  2019    12   336
 4 binance-coin 2019-12-01  15.7  15.7  15.0  15.5  2019    12   335
 5 binance-coin 2019-11-30  16.3  16.4  15.5  15.7  2019    11   334
 6 binance-coin 2019-11-29  15.7  16.3  15.6  16.3  2019    11   333
 7 binance-coin 2019-11-28  16.1  16.2  15.6  15.7  2019    11   332
 8 binance-coin 2019-11-27  15.5  16.2  14.9  16.1  2019    11   331
 9 binance-coin 2019-11-26  15.3  15.9  15.2  15.5  2019    11   330
10 binance-coin 2019-11-25  15.3  15.7  14.2  15.3  2019    11   329
# ... with 2,802 more rows

However, note that by default the date column is runed into type character.

The {rio} Package

The import() function from the {rio} package allows to load all kind of data formats:

#install.packages("rio")

data <- rio::import(here::here("data", "crypto_cleaned.csv"))

head(data) ## use just head because the output is very long

      currency       date  open  high   low close year month yday
1 binance-coin 2019-12-04 15.35 15.69 15.01 15.28 2019    12  338
2 binance-coin 2019-12-03 15.19 15.55 15.05 15.31 2019    12  337
3 binance-coin 2019-12-02 15.51 15.71 15.15 15.19 2019    12  336
4 binance-coin 2019-12-01 15.74 15.74 15.05 15.50 2019    12  335
5 binance-coin 2019-11-30 16.26 16.37 15.54 15.72 2019    11  334
6 binance-coin 2019-11-29 15.68 16.34 15.65 16.27 2019    11  333

We can turn it into a tibble afterwards—or specify it directly when importing the data set:

data <- rio::import(here::here("data", "crypto_cleaned.csv"), setclass = "tbl")

data

# A tibble: 2,812 x 9
   currency     date        open  high   low close  year month  yday
   <chr>        <date>     <dbl> <dbl> <dbl> <dbl> <int> <int> <int>
 1 binance-coin 2019-12-04  15.4  15.7  15.0  15.3  2019    12   338
 2 binance-coin 2019-12-03  15.2  15.6  15.0  15.3  2019    12   337
 3 binance-coin 2019-12-02  15.5  15.7  15.2  15.2  2019    12   336
 4 binance-coin 2019-12-01  15.7  15.7  15.0  15.5  2019    12   335
 5 binance-coin 2019-11-30  16.3  16.4  15.5  15.7  2019    11   334
 6 binance-coin 2019-11-29  15.7  16.3  15.6  16.3  2019    11   333
 7 binance-coin 2019-11-28  16.1  16.2  15.6  15.7  2019    11   332
 8 binance-coin 2019-11-27  15.5  16.2  14.9  16.1  2019    11   331
 9 binance-coin 2019-11-26  15.3  15.9  15.2  15.5  2019    11   330
10 binance-coin 2019-11-25  15.3  15.7  14.2  15.3  2019    11   329
# ... with 2,802 more rows

You could also load, for example, JSON or Excel files with the same function:

data_json <- rio::import(here::here("data", "crypto_cleaned.json"))

data_json <- as_tibble(data_json) ## somehow `setclass` doesn't work with json

data_json

# A tibble: 2,812 x 9
   currency      date  open  high   low close  year month  yday
   <chr>        <int> <dbl> <dbl> <dbl> <dbl> <int> <int> <int>
 1 binance-coin 18234  15.4  15.7  15.0  15.3  2019    12   338
 2 binance-coin 18233  15.2  15.6  15.0  15.3  2019    12   337
 3 binance-coin 18232  15.5  15.7  15.2  15.2  2019    12   336
 4 binance-coin 18231  15.7  15.7  15.0  15.5  2019    12   335
 5 binance-coin 18230  16.3  16.4  15.5  15.7  2019    11   334
 6 binance-coin 18229  15.7  16.3  15.6  16.3  2019    11   333
 7 binance-coin 18228  16.1  16.2  15.6  15.7  2019    11   332
 8 binance-coin 18227  15.5  16.2  14.9  16.1  2019    11   331
 9 binance-coin 18226  15.3  15.9  15.2  15.5  2019    11   330
10 binance-coin 18225  15.3  15.7  14.2  15.3  2019    11   329
# ... with 2,802 more rows

data_xlsx <- rio::import(
  here::here("data", "crypto_cleaned.xlsx"),
  setclass = "tbl"
)

data_xlsx

# A tibble: 2,812 x 10
   ...1  currency    date                 open  high   low close  year
   <chr> <chr>       <dttm>              <dbl> <dbl> <dbl> <dbl> <dbl>
 1 1     binance-co~ 2019-12-04 00:00:00  15.4  15.7  15.0  15.3  2019
 2 2     binance-co~ 2019-12-03 00:00:00  15.2  15.6  15.0  15.3  2019
 3 3     binance-co~ 2019-12-02 00:00:00  15.5  15.7  15.2  15.2  2019
 4 4     binance-co~ 2019-12-01 00:00:00  15.7  15.7  15.0  15.5  2019
 5 5     binance-co~ 2019-11-30 00:00:00  16.3  16.4  15.5  15.7  2019
 6 6     binance-co~ 2019-11-29 00:00:00  15.7  16.3  15.6  16.3  2019
 7 7     binance-co~ 2019-11-28 00:00:00  16.1  16.2  15.6  15.7  2019
 8 8     binance-co~ 2019-11-27 00:00:00  15.5  16.2  14.9  16.1  2019
 9 9     binance-co~ 2019-11-26 00:00:00  15.3  15.9  15.2  15.5  2019
10 10    binance-co~ 2019-11-25 00:00:00  15.3  15.7  14.2  15.3  2019
# ... with 2,802 more rows, and 2 more variables: month <dbl>,
#   yday <dbl>

We can remove the first column by using the select() function from the {dplyr} package:

data_xlsx <- dplyr::select(data_xlsx, -1)
#data_xlsx <- dplyr::select(data_xlsx, currency:yday)

data_xlsx

# A tibble: 2,812 x 9
   currency    date                 open  high   low close  year month
   <chr>       <dttm>              <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
 1 binance-co~ 2019-12-04 00:00:00  15.4  15.7  15.0  15.3  2019    12
 2 binance-co~ 2019-12-03 00:00:00  15.2  15.6  15.0  15.3  2019    12
 3 binance-co~ 2019-12-02 00:00:00  15.5  15.7  15.2  15.2  2019    12
 4 binance-co~ 2019-12-01 00:00:00  15.7  15.7  15.0  15.5  2019    12
 5 binance-co~ 2019-11-30 00:00:00  16.3  16.4  15.5  15.7  2019    11
 6 binance-co~ 2019-11-29 00:00:00  15.7  16.3  15.6  16.3  2019    11
 7 binance-co~ 2019-11-28 00:00:00  16.1  16.2  15.6  15.7  2019    11
 8 binance-co~ 2019-11-27 00:00:00  15.5  16.2  14.9  16.1  2019    11
 9 binance-co~ 2019-11-26 00:00:00  15.3  15.9  15.2  15.5  2019    11
10 binance-co~ 2019-11-25 00:00:00  15.3  15.7  14.2  15.3  2019    11
# ... with 2,802 more rows, and 1 more variable: yday <dbl>

Aesthetics: aes()

Some prefer to place the aes() outside the ggplot() call:

ggplot(data) +
  aes(x = date, y = close)

Coordinate Systems: coord_*()

The coordinate system maps the two position to a 2d position on the plot:

ggplot(data, aes(x = date, y = close, 
                 color = currency)) +
  geom_line() +
  geom_point() +
  scale_x_date() + 
  scale_y_continuous() + 
  scale_color_discrete() +
  coord_cartesian()

ggplot(data, aes(x = date, y = close, 
                 color = currency)) +
  geom_line() +
  geom_point() +
  scale_x_date() + 
  scale_y_continuous() + 
  scale_color_discrete() +
  coord_polar()

Changing the limits on the coordinate system allows to zoom in:

ggplot(data, aes(x = date, y = close, 
                 color = currency)) +
  geom_line() +
  geom_point() +
  scale_x_date() + 
  scale_y_continuous() + 
  scale_color_discrete() +
  coord_cartesian(
    xlim = c(as.Date("2018-11-01"),
             as.Date("2019-11-01")),
    ylim = c(NA, 100)
  )

Session Info

Expand for details

[1] "2022-06-25 13:48:26 CEST"

Local:    main C:/Users/DataVizard/Google Drive/Work/DataViz/Teaching/2022_OReilly_Trainings/hands-on-ggplot2-training
Remote:   main @ origin (https://github.com/z3tt/hands-on-ggplot2-training.git)
Head:     [1cc6183] 2022-02-18: fix typo

R version 4.1.0 (2021-05-18)
Platform: x86_64-w64-mingw32/x64 (64-bit)
Running under: Windows 10 x64 (build 19043)

Matrix products: default

locale:
[1] LC_COLLATE=German_Germany.1252  LC_CTYPE=German_Germany.1252   
[3] LC_MONETARY=German_Germany.1252 LC_NUMERIC=C                   
[5] LC_TIME=German_Germany.1252    
system code page: 65001

attached base packages:
[1] stats     graphics  grDevices utils     datasets  methods  
[7] base     

other attached packages:
[1] forcats_0.5.1   stringr_1.4.0   dplyr_1.0.7     purrr_0.3.4    
[5] readr_2.0.2     tidyr_1.1.4     tibble_3.1.6    ggplot2_3.3.5  
[9] tidyverse_1.3.1

loaded via a namespace (and not attached):
 [1] httr_1.4.2        sass_0.4.0        bit64_4.0.5      
 [4] vroom_1.5.5       jsonlite_1.7.2    here_1.0.1       
 [7] modelr_0.1.8      bslib_0.3.1       assertthat_0.2.1 
[10] highr_0.9         cellranger_1.1.0  yaml_2.2.1       
[13] pillar_1.6.4      backports_1.2.1   glue_1.4.2       
[16] digest_0.6.29     rvest_1.0.2       colorspace_2.0-2 
[19] htmltools_0.5.2   pkgconfig_2.0.3   broom_0.8.0      
[22] haven_2.4.3       scales_1.1.1      openxlsx_4.2.5   
[25] rio_0.5.29        distill_1.3       tzdb_0.1.2       
[28] downlit_0.4.0     git2r_0.29.0      generics_0.1.1   
[31] farver_2.1.0      ellipsis_0.3.2    cachem_1.0.6     
[34] withr_2.4.3       cli_3.1.0         magrittr_2.0.1   
[37] crayon_1.4.2      readxl_1.3.1      memoise_2.0.1    
[40] evaluate_0.15     fs_1.5.0          fansi_0.5.0      
[43] foreign_0.8-81    xml2_1.3.2        textshaping_0.3.6
[46] data.table_1.14.2 tools_4.1.0       hms_1.1.1        
[49] lifecycle_1.0.1   munsell_0.5.0     reprex_2.0.1     
[52] zip_2.2.0         compiler_4.1.0    jquerylib_0.1.4  
[55] systemfonts_1.0.3 rlang_1.0.2       grid_4.1.0       
[58] rstudioapi_0.13   labeling_0.4.2    rmarkdown_2.11   
[61] gtable_0.3.0      DBI_1.1.2         curl_4.3.2       
[64] R6_2.5.1          lubridate_1.8.0   knitr_1.39       
[67] fastmap_1.1.0     bit_4.0.4         utf8_1.2.2       
[70] rprojroot_2.0.2   ragg_1.1.3        stringi_1.7.5    
[73] parallel_4.1.0    Rcpp_1.0.7        vctrs_0.3.8      
[76] dbplyr_2.1.1      tidyselect_1.1.1  xfun_0.31

Segment 1: The Basics of ggplot2

The {ggplot2} Package

Advantages of {ggplot2}

The Setup

The Course Material

The Package

The Data Set

Import the Data

Short Explanation of ::

Exercise 1

The Grammar of Graphics

The Structure of ggplot2

Data: ggplot(data)

Aesthetics: aes()

Layers: geom_*() and stat_*()

Aesthetics: aes() (again)

Important Things First: Change the Default Theme

Layers: geom_*() and stat_*() (repeated)

Exercise 2

Saving a ggplot

Resources

Appendix

Import the Data

The {here} Package

Base R

The {rio} Package

Aesthetics: aes()

Coordinate Systems: coord_*()

Session Info

The `{ggplot2}` Package

Advantages of `{ggplot2}`

Layers: geom_() and stat_()

Layers: geom_() and stat_() (repeated)

Saving a `ggplot`

The `{here}` Package